Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH v1 00/11] KVM: s390: pv: implement lazy destroy
To:     Claudio Imbrenda <imbrenda@linux.ibm.com>,
        Cornelia Huck <cohuck@redhat.com>
Cc:     kvm@vger.kernel.org, frankja@linux.ibm.com, thuth@redhat.com,
        pasic@linux.ibm.com, david@redhat.com, linux-s390@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <20210517200758.22593-1-imbrenda@linux.ibm.com>
 <20210518170537.58b32ffe.cohuck@redhat.com> <20210518173624.13d043e3@ibm-vm>
From:   Christian Borntraeger <borntraeger@de.ibm.com>
Message-ID: <225fe3ec-f2e9-6c76-97e1-b252fe3326b3@de.ibm.com>
Date:   Tue, 18 May 2021 17:45:18 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.10.0
MIME-Version: 1.0
In-Reply-To: <20210518173624.13d043e3@ibm-vm>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 18.05.21 17:36, Claudio Imbrenda wrote:
> On Tue, 18 May 2021 17:05:37 +0200
> Cornelia Huck <cohuck@redhat.com> wrote:
> 
>> On Mon, 17 May 2021 22:07:47 +0200
>> Claudio Imbrenda <imbrenda@linux.ibm.com> wrote:
>>
>>> Previously, when a protected VM was rebooted or when it was shut
>>> down, its memory was made unprotected, and then the protected VM
>>> itself was destroyed. Looping over the whole address space can take
>>> some time, considering the overhead of the various Ultravisor Calls
>>> (UVCs).  This means that a reboot or a shutdown would take a
>>> potentially long amount of time, depending on the amount of used
>>> memory.
>>>
>>> This patchseries implements a deferred destroy mechanism for
>>> protected guests. When a protected guest is destroyed, its memory
>>> is cleared in background, allowing the guest to restart or
>>> terminate significantly faster than before.
>>>
>>> There are 2 possibilities when a protected VM is torn down:
>>> * it still has an address space associated (reboot case)
>>> * it does not have an address space anymore (shutdown case)
>>>
>>> For the reboot case, the reference count of the mm is increased, and
>>> then a background thread is started to clean up. Once the thread
>>> went through the whole address space, the protected VM is actually
>>> destroyed.
>>>
>>> For the shutdown case, a list of pages to be destroyed is formed
>>> when the mm is torn down. Instead of just unmapping the pages when
>>> the address space is being torn down, they are also set aside.
>>> Later when KVM cleans up the VM, a thread is started to clean up
>>> the pages from the list.
>>
>> Just to make sure, 'clean up' includes doing uv calls?
> 
> yes
> 
>>>
>>> This means that the same address space can have memory belonging to
>>> more than one protected guest, although only one will be running,
>>> the others will in fact not even have any CPUs.
>>
>> Are those set-aside-but-not-yet-cleaned-up pages still possibly
>> accessible in any way? I would assume that they only belong to the
> 
> in case of reboot: yes, they are still in the address space of the
> guest, and can be swapped if needed
> 
>> 'zombie' guests, and any new or rebooted guest is a new entity that
>> needs to get new pages?
> 
> the rebooted guest (normal or secure) will re-use the same pages of the
> old guest (before or after cleanup, which is the reason of patches 3
> and 4)
> 
> the KVM guest is not affected in case of reboot, so the userspace
> address space is not touched.
> 
>> Can too many not-yet-cleaned-up pages lead to a (temporary) memory
>> exhaustion?
> 
> in case of reboot, not much; the pages were in use are still in use
> after the reboot, and they can be swapped.
> 
> in case of a shutdown, yes, because the pages are really taken aside
> and cleared/destroyed in background. they cannot be swapped. they are
> freed immediately as they are processed, to try to mitigate memory
> exhaustion scenarios.
> 
> in the end, this patchseries is a tradeoff between speed and memory
> consumption. the memory needs to be cleared up at some point, and that
> requires time.
> 
> in cases where this might be an issue, I introduced a new KVM flag to
> disable lazy destroy (patch 10)

Maybe we could piggy-back on the OOM-kill notifier and then fall back to
synchronous freeing for some pages?